Great Circle Associates Majordomo-Workers
(June 1995)
 

Indexed By Date: [Previous] [Next] Indexed By Thread: [Previous] [Next]

Subject: majordomo 1.93 bug?
From: "Joseph Kwan" <jkwan @ ampersand . Jpl . Nasa . Gov>
Date: Tue, 20 Jun 1995 16:46:59 -0700
To: Majordomo-Workers @ greatcircle . com
Cc: jkwan @ ampersand . Jpl . Nasa . Gov


I was having a problem with majordomo 1.92 where I'd send majordomo a
request ("who" or "list") and it'd sit there for some time (5 mins I
think), then it would finally respond to the request.  I grabbed 1.93
and it did the same thing.  Upon further debugging (trace -p PID),
I discovered that it was looping trying to stat "shlock.PID" in the
log file directory that it did not have permission.

The difference between 1.92 and 1.93 is that in 1.93, it chews up
memory and eventually the process gets an "Out of memory" error (one
of the return mail messages I get from the system is the 'out of
memory' error message).  Both version do return the correct results,
but one eats up all the memory.  This could be a perl problem (I'm
running perl 5 with both versions), but I'm not 100% sure.

fyi.  (of course, everything works fine if majordomo has the correct
permissions to write into the log directory; I just discovered that
the permissions on the log directory had changed after I noticed
the delays with majordomo).

Joe

Joseph Kwan, Sr. Network/Systems Analyst
Radar Science and Engineering (Section 334)
Jet Propulsion Laboratory / MS 300-235
Internet: jkwan@radar-sci.jpl.nasa.gov / X.500: Joe.Kwan@jpl.nasa.gov


Indexed By Date Previous: My problem with get and locks
From: Jeff Okamoto <okamoto@hpcc123.corp.hp.com>
Next: Minor bug in 1.93
From: dwolfe@risc.sps.mot.com (Dave Wolfe)
Indexed By Thread Previous: My problem with get and locks
From: Jeff Okamoto <okamoto@hpcc123.corp.hp.com>
Next: Minor bug in 1.93
From: dwolfe@risc.sps.mot.com (Dave Wolfe)

Google
 
Search Internet Search www.greatcircle.com